| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| SVRCONN channel -UOW state none & Channel substate Recei |
« View previous topic :: View next topic » |
| Author |
Message
|
| rickwatsonb |
 Posted: Thu Jul 09, 2009 12:24 pm Post subject: SVRCONN channel -UOW state none & Channel substate Recei |
 |
|
Voyager
Joined: 15 Aug 2006
Posts: 87
Location: USA: Mid-West
|
Hi,
We have a MQ Client/MQ server architecture that uses a SVRCONN channel, with no connection pooling, and all on SUN servers. Some of the connections show an MQ Explorer Application Connection Status, UOW state of ‘none’, and a Channel Status, channel substate of ‘receiving’. Unlike the other connections, there is no queue associated with the Application Connection status with a UOW state of ‘none’. (For simplicity, I refer to these connections as ‘unknown’ connections below.) These states are abnormal to the rest of the connections to the queue manger; all of the other connections show a state of ‘active’ and an ‘In MQGET’ state respectively.
What do the UOW state of ‘none’, and a channel substate of ‘receiving’ imply?
Is it correct to classify these channels as orphaned SVRCONN channels? If so, can the cause be due to dropped TCP connections, and/or missing disconnect and close procedures? Are there other causes? Who should I hound on the project for data/answers?
Extraneous detail:
In general, I am expecting approx. 483 connections to the queue manager relative to the application channel. But, on a daily basis, we typically have 594 connections (count via netstat and port number).
In more detail, a total of 6 application servers provide a total of 23 instances of the application (5 servers have 4 instances of the application and the sixth has 3). There are 21 application queues, thus yielding the 483 connections amongst the 23 application instances.
Specifically, there are 5 servers that have 16 unknown connections each, and one server has 13 unknown connections (server with 3 instances), this gives a total of 96 unknown connections. (There are 15 connections related to perhaps some other problem (last message and start date of May 25)). (594 – ((5 x 16 + 13) - 15) = 483 and (23 x 21 = 483).
I would appreciate any help you can offer.
Thank you. |
|
| Back to top |
|
 |
| rickwatsonb |
| Posted: Thu Jul 09, 2009 12:33 pm Post subject: |
|
|
Voyager
Joined: 15 Aug 2006
Posts: 87
Location: USA: Mid-West
|
Hi,
Adding detail to my own post:
MQ 6.0.2.0 server on Solaris 10
TCP KeepAlive = 2 hours
Channel:
CHANNEL(APP.CHANNEL) CHLTYPE(SVRCONN)
ALTDATE(2008-05-11) ALTTIME(03.37.26)
COMPHDR(NONE) COMPMSG(NONE)
DESCR( ) HBINT(300)
KAINT(AUTO) MAXMSGL(4194304)
MCAUSER(<someid>) MONCHL(QMGR)
RCVDATA( ) RCVEXIT( )
SCYDATA( ) SCYEXIT( )
SENDDATA( ) SENDEXIT( )
SSLCAUTH(REQUIRED) SSLCIPH( )
SSLPEER( ) TRPTYPE(TCP) |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Thu Jul 09, 2009 1:27 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
Huh. "None" is not listed as something you would see.
http://publib.boulder.ibm.com/infocenter/wmqv6/v6r0/topic/com.ibm.mq.csqzaj.doc/sc11860_.htm
I looked at a couple of my QMs that have > 1000 connections and none have None.
You say KeepAlive is set to 2 hours, but does the qm.ini file have an entry that tells the QM to actually use the O/S'es KeepAlive value?
We set ours to 15 minutes, 2 hours is a lot.
If the QM is in fact using Keep Alive, and its 2 hours, do any of these "None" channels have a start time > 2 hours?
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| rickwatsonb |
| Posted: Fri Jul 10, 2009 6:34 am Post subject: |
|
|
Voyager
Joined: 15 Aug 2006
Posts: 87
Location: USA: Mid-West
|
Hi,
Thank you for the reply PeterPotkay.
PeterPotkay wrote:
[Huh. "None" is not listed as something you would see.
http://publib.boulder.ibm.com/infocenter/wmqv6/v6r0/topic/com.ibm.mq.csqzaj.doc/sc11860_.htm ]
Sorry, but I miss seeing the purpose of the link above relative to "none" and the UOW state. The contact admin pages have the following for Application Connection Properties, UOW state:
(Excerpt) Unit of work state - The state of the unit of work. None means that there is no unit of work; Active means that the unit of work is active;... (I did not see an exact definition for a "unit of work' but it appears to be related to a "transaction associated with the connection". A detailed understanding of what "none" means is lacking here...)
In reference to TCP KeepAlive:
No, we are not using it in the qm.ini file. I will work on implementing it in a lower environment and see if this resolves the unkown channels.
From reading many previous posts, the application code, which was written when 5.3 was out, will also need to be critiqued for usage of disconnect and close commands. That will be harder to get at...
Thanks for your help; having a conversation with myself gets a bit old sometimes.
[img]  [/img] |
|
| Back to top |
|
|
| PeterPotkay |
| Posted: Fri Jul 10, 2009 11:37 am Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
In the link I posted, scroll down to the SUBSTATE section and look at the various possibilities - "NONE" is not listed.
_________________
Peter Potkay
Keep Calm and MQ On |
|
| Back to top |
|
|
| rickwatsonb |
| Posted: Fri Jul 10, 2009 12:17 pm Post subject: |
|
|
Voyager
Joined: 15 Aug 2006
Posts: 87
Location: USA: Mid-West
|
Hi,
Thank you for your reply PeterPotkay.
In my first post, I referred to data from two different MQ Explorer options:
(1) Application Connection Status, UOW state of ‘none’ (right click queue manager)
(2) Channel Status, channel substate of ‘receiving’ (right click channel).
The value of 'none' is relative to the Application Connection Status, not the substate channel status. I hope this makes more sense.
Testing:
This afternoon I edited the qm.ini on a dev queue manager, recycled it. The app server was also recycled so it would know to connect again (a quirk we have). Currently there are network issues and the app has not re-connected, so I will post results tomorrow perhaps.
Thank you. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri Jul 10, 2009 10:09 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| rickwatsonb wrote: |
Hi,
Thank you for your reply PeterPotkay.
In my first post, I referred to data from two different MQ Explorer options:
(1) Application Connection Status, UOW state of ‘none’ (right click queue manager)
(2) Channel Status, channel substate of ‘receiving’ (right click channel).
The value of 'none' is relative to the Application Connection Status, not the substate channel status. I hope this makes more sense.
Testing:
This afternoon I edited the qm.ini on a dev queue manager, recycled it. The app server was also recycled so it would know to connect again (a quirk we have). Currently there are network issues and the app has not re-connected, so I will post results tomorrow perhaps.
Thank you. |
I think your point 1 might refer to whether the application is using syncpoint or not... 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
